MiDaS V2 INT8

Cloud CPUAWS Graviton G4

An INT8-quantized monocular depth estimation model that predicts relative inverse-depth maps from single RGB images. It uses a ResNet-based backbone and produces dense depth outputs suitable for applications such as AR/VR, robotics, novel-view synthesis, and image editing. Inputs are resized RGB images, and outputs are relative depth maps requiring calibration for metric depth.View on Hugging Face

Accuracyvs base model

AbsRel0.12↑0

RMSE0.57 m↑0.01 m

δ1 < 1.2586.51%

About this version

Original Model:: midas_v2 by Intel — midas_v2
Optimization/conversion:: Arm-Optimized version for execution on Arm-based platforms.
Converted/optimized by:: Arm
License:: The Original Model and the Optimized Model are subject to MIT.

This repository contains a converted or optimized version of the Original Model (the “Optimized Model”). The Original Model has been converted or optimized as described above for execution on Arm-based platforms.

No retraining or fine-tuning of the Original Model was performed as part of the conversion or optimization. The conversion or optimization was not intended to change the Original Model’s behavior or intended use.

Original Model and Documentation

For information about the Original Model, including its development, training data, intended uses, limitations and other relevant information, please refer to the Original Model repository. Information in that repository was provided by the original developer or other third parties and, unless expressly stated otherwise, has not been independently verified by Arm.

Licenses and Third-Party Terms

Use of the Original Model and the Optimized Model is subject to the applicable licenses, usage restrictions and other terms identified above and in the relevant repositories. Publication of the Optimized Model does not grant any rights beyond those provided under the applicable license terms.

You are responsible for reviewing those terms and ensuring that your use of the Original Model and the Optimized Model is permitted.

Purpose of this Release

The Optimized Model is provided as a reference implementation to demonstrate and evaluate execution and performance on Arm-based systems. It is not a production-ready or supported solution.

About this version

Original Model:: midas_v2 by Intel — midas_v2
Optimization/conversion:: Arm-Optimized version for execution on Arm-based platforms.
Converted/optimized by:: Arm
License:: The Original Model and the Optimized Model are subject to MIT.

Original Model and Documentation

Licenses and Third-Party Terms

You are responsible for reviewing those terms and ensuring that your use of the Original Model and the Optimized Model is permitted.

Purpose of this Release

The Optimized Model is provided as a reference implementation to demonstrate and evaluate execution and performance on Arm-based systems. It is not a production-ready or supported solution.

Metric	Resultsvs baseline	Change
End to end latency p50	104.91ms454.97ms	−350.06 ms
End to end latency p90	105.03ms455.15ms	−350.11 ms
End to end latency p99	105.19ms458.04ms	−352.86 ms
Peak memory	209.50MB792.00MB	−582.50 MB
Time to first inference	131.57ms482.22ms	−350.64 ms
Frames per second	9.53fps2.20fps	+7.33 fps
AbsRel	0.120.11	0.00
RMSE	0.57m0.57m	+0.01 m
δ1 < 1.25	86.51%87.02%	−0.51%

Dataset	nyu_depth_v2	nyu_depth_v2
Sample count	654	654
Batch size	1	1
Input resolution	384x288	384x288
Runs	50	50
Warmup runs	10	10

MiDaS V2 INT8

MiDaS V2 INT8

Model Evaluation

Configuration

Comparability